Re: Deja vu: panic in hdaa_coonfigure() for i386, but not amd64 -- again

On 05/16/15 15:37, David Wolfskill wrote:
> On Sat, May 16, 2015 at 03:27:57PM +0200, Hans Petter Selasky wrote:
>> ...
>>> After reverting both 282650 & 282651, the resulting kernel built.
>>>
>>
>> Hi,
>>
>> I'm seeing two of my commits mentioned here, r282650 & r282651. Are
>> these what is causing the problem? You are sure the built the kernel and
>> all modules from clean, hence some structures are changed, panic would
>> be expected if you don't.
>> ....
>

Hi,

> I'll admit that the vast bulk of my builds are done with -DNOCLEAN -- I
> actually want to use my laptop for things other than rebuilding FreeBSD
> (4x/day) and rebuilding updated installed ports (2x/day).

Then it will possibly trigger bad, because AFMT_CHANNELS() returns part 
of the old EXT channels, which is used in the HDAA driver. If you want 
to save time, you need to rebuild all of "modules/sound" only, w/o 
-DNO_CLEAN and doing "clean" first. The rest of the kernel is fine with 
-DNO_CLEAN.

> After I finish the build I'm presently doing, I will try re-doing
> r282650 & r282651, then rebuilding cleanly.
>
> I noted that r282650 included a bump to param.h to force all modules to
> be rebuilt.
>
> I also note that the panic was only seen in head/i386 -- head/amd64 on
> the same hardware, using the same procedures, sources updated from the
> same local private mirror of the SVN repo, using a copy of the same
> "CANARY" kernel config file, had no issue.

Send output from bootverbose! Let's hunt this bug down :-)

--HPS